In the 60’s of the last century, Mullins and Sekerka had a good conclusion for crystal growth with no convection in the fluids. But in the process of crystal growth, there exists the convection in the fluids, and the velocity of the melted fluid should satisfy both the continuity equation and the momentum equation. By using Mullins-Sekerka method, as the influences of the convection included, in the first-order approximation it would be found out that the velocity of the melted fluid on solid-liquid interface front should be satisfied a harmonic function. But generally that is not the case. It shows that in the case the influences of the convection is included, the Mullins-Sekerka method should be modified.